OpenAI adds Sketch to ChatGPT for turning doodles into images

In short: OpenAI has updated ChatGPT’s image tool and added Sketch, a feature that turns your quick doodles into more detailed images.

What happened

OpenAI announced “ChatGPT Images 2.5,” an update to the image making tools inside ChatGPT. The company also added a new feature called Sketch.

Sketch lets you draw a simple picture directly inside ChatGPT and use it as part of your request. Think of it like giving directions with both words and a rough map (your doodle) so ChatGPT has a clearer idea of what you mean.

To use it, you type @Sketch in the chat box. A drawing window appears, and you can sketch with your mouse or finger. The Verge reporter Jay Peters tested it by drawing a rough cat and asking ChatGPT to turn it into a realistic photo.

OpenAI also says you can now leave comments on specific parts of an image to ask for changes. For example, Peters used this to change the cat’s eye color.

Along with Sketch, OpenAI says Images 2.5 makes images with more natural lighting and richer textures. The company also says it follows editing instructions better over multiple messages, and it can generate images faster, with latency (waiting time) reduced by up to 50 percent compared with Images 2.0.

ChatGPT Images 2.5 is available now for ChatGPT, ChatGPT Work, and Codex users on desktop, mobile, and web.

Why it matters

Many people struggle to describe what they want in words. Sketch gives them a simple way to show it instead, which could make image editing and creation faster and less frustrating for everyday use.

Source: The Verge AI